Modeling complexity in nature: A Bayesian hierarchical framework

Bayesian statistics provides a flexible framework for understanding the complexity of wildlife populations by modeling ecological processes. In this work, I developed and applied Bayesian hierarchical models to key questions in population dynamics. First, I connected classical open population models with genetic information to model population dynamics close to extinction, aiming to quantify parameters such as population size and survival probabilities. Second, I integrated different data sources to model species distribution across large spatial and temporal scales, combining opportunistic observations with structured surveys to produce spatial distribution maps and abundance estimates. Third, I implemented methodologies to infer habitat preferences and associations, linking environmental variables to demographic rates and the distribution of individuals. By taking advantage of the Bayesian hierarchical structure, it is possible to, firstly, diagnose challenges in the modeling of small populations, including weak identifiability of covariate effects, secondly, provide a scalable approach for estimating abundance and distribution over broad spatio-temporal extents and thirdly, quantify the effect of environmental variables in the distribution of wild individuals. The resulting analyses were translated into conservation actions, aiming to support the decision-making process and improve wildlife management.